Simplify proc macro bridge state

[joboet]

Currently, proc_macro uses a ScopedCell to store the client-side proc-macro bridge. Unfortunately, this requires the Bridge, which has non-negligible size, to be copied out and back again on every access. In some cases, the optimizer might be able to elide these copies, but in general, this is suboptimal.

This PR removes ScopedCell and employs a similar trick as in scoped_tls, meaning that the only thing stored in TLS is a pointer to the state, which now is a RefCell. Access to the pointer is then scoped so that it is always within the lifetime of the reference to the state. Unfortunately, scoped_tls requires the referenced type to be 'static, which Bridge is not, therefore we cannot simply copy that macro but have to reimplement its TLS trick.

This removes the #[forbid(unsafe_code)] on the client module so that we do not have to export Bridge, which currently is private, to the whole crate. I can change that, if necessary.

[rustbot]

r? @petrochenkov

rustbot has assigned @petrochenkov.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[joboet]

@bors try @rust-timer queue

[bors]

⌛ Trying commit ac770f7 with merge b9b0f25...

[bors]

☀️ Try build successful - checks-actions
Build commit: b9b0f25 (b9b0f257f9615732562e3a469bfef0ca3379a935)

[rust-timer]

Finished benchmarking commit (b9b0f25): comparison URL.

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.5%	[0.4%, 0.6%]	2
Improvements ✅ (primary)	-0.4%	[-0.5%, -0.3%]	3
Improvements ✅ (secondary)	-1.2%	[-1.9%, -0.5%]	6
All ❌✅ (primary)	-0.4%	[-0.5%, -0.3%]	3

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.1%	[-2.1%, -2.1%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.1%	[-2.1%, -2.1%]	1

Cycles

This benchmark run did not return any relevant results for this metric.

Binary size

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	2.3%	[2.2%, 2.5%]	4
Regressions ❌ (secondary)	5.9%	[5.9%, 5.9%]	3
Improvements ✅ (primary)	-0.6%	[-0.6%, -0.6%]	3
Improvements ✅ (secondary)	-1.9%	[-1.9%, -1.9%]	3
All ❌✅ (primary)	1.0%	[-0.6%, 2.5%]	7

Bootstrap: 669.588s -> 670.102s (0.08%)
Artifact size: 315.05 MiB -> 315.01 MiB (-0.02%)

[joboet]

Okay, that's quite good, especially the token-stream-stress. I'll see if integrating the RefCell into the pointer helps tt-muncher.

@rustbot author

[joboet]

Yeah, that's worse. Let's just use RefCell.
@rustbot ready

[petrochenkov]

@bors r+

[bors]

📌 Commit ac770f7 has been approved by petrochenkov

It is now in the queue for this repository.

[bors]

⌛ Testing commit ac770f7 with merge 536606b...

[bors]

☀️ Test successful - checks-actions
Approved by: petrochenkov
Pushing 536606b to master...

[rust-log-analyzer]

A job failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

  PR_MESSAGE: Automation to keep dependencies in `Cargo.lock` current.
following is the output from `cargo update`:
  COMMIT_MESSAGE: cargo update 
##[endgroup]
Starting download for Cargo-lock
##[error]Unable to find any artifacts for the associated workflow

[rust-timer]

Finished benchmarking commit (536606b): comparison URL.

Overall result: ✅ improvements - no action needed

@rustbot label: -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.3%	[0.3%, 0.3%]	1
Improvements ✅ (primary)	-0.3%	[-0.4%, -0.3%]	2
Improvements ✅ (secondary)	-1.2%	[-1.9%, -0.4%]	6
All ❌✅ (primary)	-0.3%	[-0.4%, -0.3%]	2

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.6%	[0.5%, 3.4%]	13
Regressions ❌ (secondary)	3.1%	[2.6%, 4.0%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.6%	[0.5%, 3.4%]	13

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.1%	[0.5%, 3.2%]	11
Regressions ❌ (secondary)	2.4%	[2.4%, 2.4%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.9%	[-3.4%, -2.4%]	3
All ❌✅ (primary)	1.1%	[0.5%, 3.2%]	11

Binary size

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	2.3%	[2.2%, 2.5%]	4
Regressions ❌ (secondary)	5.9%	[5.9%, 5.9%]	3
Improvements ✅ (primary)	-0.6%	[-0.6%, -0.6%]	3
Improvements ✅ (secondary)	-1.9%	[-1.9%, -1.9%]	3
All ❌✅ (primary)	1.0%	[-0.6%, 2.5%]	7

Bootstrap: 672.189s -> 671.741s (-0.07%)
Artifact size: 315.72 MiB -> 315.68 MiB (-0.01%)

[rust-log-analyzer]

A job failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

  PR_MESSAGE: Automation to keep dependencies in `Cargo.lock` current.
following is the output from `cargo update`:
  COMMIT_MESSAGE: cargo update 
##[endgroup]
Starting download for Cargo-lock
##[error]Unable to find any artifacts for the associated workflow

[therealprof]

Uhm, why does this have an impact on the binary size... This should not result in different code, should it?

[joboet]

All the changes are in proc-macro crates, so this makes sense. I'd wager that this change led to more proc_macro code being inlined, hence the increase. Given that this is only noticeable on development machines, I don't think this is bad, especially considering the performance improvements we get in return.